teradataml 17.20.0.7__py3-none-any.whl → 20.0.0.0__py3-none-any.whl

teradataml/catalog/model_cataloging.py

@@ -1,980 +0,0 @@
-# -*- coding: utf-8 -*-
-"""
-Unpublished work.
-Copyright (c) 2020 by Teradata Corporation. All rights reserved.
-TERADATA CORPORATION CONFIDENTIAL AND TRADE SECRET
-
-Primary Owner: Rohit.Khurd@teradata.com
-Secondary Owner:
-
-teradataml Model Cataloging
--------------------------------
-teradataml Model Cataloging functions provide interface to manage user created models
-generated by teradataml Analytic functions, and retrieve saved models to use with teradataml.
-"""
-
-import pandas as pd
-import json
-import warnings
-
-from teradataml.common.constants import ModelCatalogingConstants as mac
-from collections import OrderedDict
-from teradataml.common.exceptions import TeradataMlException
-from teradataml.common.garbagecollector import GarbageCollector
-from teradataml.common.messagecodes import MessageCodes
-from teradataml.common.messages import Messages
-from teradataml.context.context import get_connection
-from teradataml.common.utils import UtilFuncs, function_deprecation
-from teradataml.dbutils.dbutils import _execute_transaction
-from teradataml.options.display import *
-from teradataml.utils.validators import _Validators
-from teradataml.catalog.function_argument_mapper import _argument_mapper
-from teradatasql import OperationalError as SqlOperationalError
-from teradataml.dataframe.dataframe import DataFrame, in_schema
-
-from teradataml.catalog.model_cataloging_utils import __check_if_model_cataloging_tables_exists,\
-    __check_if_model_exists, __get_model_engine, __get_current_user, __get_model_inputs_outputs,\
-    __get_model_parameters, __get_like_filter_expression_on_col, __retrieve_model_client_engine_algorithm,\
-    __retrieve_model_attributes, __retrieve_model_outputs, __get_tables_for_model, __delete_model_tableview,\
-    __retrieve_argument_and_output_map, __get_wrapper_class, __get_model_access, __retrieve_model_inputs
-
-
-__awu = _Validators()
-__gc = GarbageCollector()
-
-
-# TODO: Support for type argument in all MC API's.
-#       This should be done when DLEngine support is added.
-#       type - Required Argument. Indicates the type of the model. ML Engine or DL Engine.
-
-# TODO: Notes for save_model()
-#   1. TODO: Add support for overwrite = True
-#
-@function_deprecation("future")
-def save_model(model,
-               name,
-               description,
-               model_project=None,
-               entity_target=None,
-               performance_metrics=None):
-    """
-    DESCRIPTION:
-        Function to save a teradataml Analytic Function model in Teradata Vantage.
-
-    PARAMETERS:
-        model:
-            Required Argument.
-            Specifies the teradataml analytic function model to be saved.
-            Types: teradataml analytic function object.
-
-        name:
-            Required Argument.
-            Specifies the unique name to identify the saved model.
-            The maximum length of the name is 128 characters.
-            Types: str
-
-        description:
-            Required Argument.
-            Specifies a note describing the model to be saved.
-            The maximum length of the description is 1024 characters.
-            Types: str
-
-        model_project:
-            Optional Argument.
-            Specifies the project that the model is associated with.
-            The maximum length of the model_project is 128 characters.
-            Types: str
-            Default Value: None
-
-        entity_target:
-            Optional Argument.
-            Specifies a group or team that the model is associated with.
-            The maximum length of the entity_target is 128 characters.
-            Types: str
-            Default Value: None
-
-        performance_metrics:
-            Optional Argument.
-            Specifies the performance metrics for the model.
-            performance_metrics must be a dictionary of the following form:
-                { "<metric>" : { "measure" : <value> }, ... }
-            For example:
-                { "AUC" : { "measure" : 0.5 }, ... }
-            The value should be of type float.
-            Types: dict
-            Default Value: None
-
-    RETURNS:
-        None.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforest", ["housing_train"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-
-        # This example uses home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, False, (str), True])
-    __arg_info_matrix.append(["description", description, False, (str), True])
-    __arg_info_matrix.append(["model_project", model_project, True, (str), True])
-    __arg_info_matrix.append(["entity_target", entity_target, True, (str), True])
-    __arg_info_matrix.append(["performance_metrics", performance_metrics, True, (dict)])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    # Check length of the model name provided by user.
-    __awu._validate_str_arg_length('name', name, 'LE', 128)
-    # Check length of the model description provided by user.
-    __awu._validate_str_arg_length('description', description, 'LE', 1024)
-    # Check length of the model project name provided by user, if not None.
-    if model_project:
-        __awu._validate_str_arg_length('model_project', model_project, 'LE', 128)
-    # Check length of the model entity target name provided by user, if not None.
-    if entity_target:
-        __awu._validate_str_arg_length('entity_target', entity_target, 'LE', 128)
-
-    # Check type of model argument. If it's not MLE or SQLE model, raise the error.
-    if not ("teradataml.analytics.mle." in str(type(model)) or "teradataml.analytics.sqle." in str(type(model))):
-        raise TypeError(Messages.get_message(MessageCodes.UNSUPPORTED_DATATYPE, ["model"],
-                                             ["teradataml Analytic Function Object"]))
-
-    # Check if model._sql_specific_attributes is None
-    if model._get_sql_specific_attributes() is None:
-        # This means that the model is a retrieved model, and we should raise an exception here
-        raise TeradataMlException(Messages.get_message(MessageCodes.CANNOT_SAVE_RETRIEVED_MODEL),
-                                  MessageCodes.CANNOT_SAVE_RETRIEVED_MODEL)
-
-    # Check if model with same name exists or not, raise error if it exists.
-    __check_if_model_exists(name, raise_error_if_exists=True)
-
-    # Validations are complete. Let's move forward with model saving.
-    # We first gather the model related information.
-    generating_engine = __get_model_engine(model)
-    generating_client = mac.MODEL_TDML.value
-    algorithm = model._get_algorithm_name().lower()
-    build_time = model.get_build_time()
-
-    # Get the current user name.
-    conn = get_connection()
-    created_by = __get_current_user(conn)
-    status = mac.DEFAULT_SAVE_STATUS.value
-    access = mac.DEFAULT_SAVE_ACCESS.value
-    location = mac.MODEL_ENGINE_ADVSQL.value
-
-    # Let's get the attributes, input and output related information.
-    function_arg_map = _argument_mapper._get_function_map(engine=generating_engine,
-                                                          function_name=algorithm)
-    vantage_version = _argument_mapper._vantage_version
-
-    training_data, training_objects, remove_tables_entries_from_gc = __get_model_inputs_outputs(model,
-                                                                                                function_arg_map)
-    training_data = json.dumps(training_data)
-    training_objects = json.dumps(training_objects)
-
-    algorithm_parameters = json.dumps(__get_model_parameters(model, function_arg_map))
-    # Escape single quotes in values
-    algorithm_parameters = algorithm_parameters.replace("'", "''")
-
-    target_column = model.get_target_column()
-    prediction_type = model.get_prediction_type()
-
-    if performance_metrics:
-        performance_metrics = json.dumps(performance_metrics)
-    else:
-        performance_metrics = None
-
-    # Save the model based on the new details provided.
-    # Call the Stored Procedure to save the model.
-    # But first, find the transaction mode of the session to decide which Stored Procedure to call
-    save_sp = '{}{}'.format(mac.SAVE_MODEL.value, 'T')
-    transaction_mode = conn.dialect.get_transaction_mode(conn)
-    if transaction_mode and transaction_mode.lower() == 'a':
-        save_sp = '{}{}'.format(mac.SAVE_MODEL.value, 'A')
-
-    # Form the stored procedure call
-    sp_call = "CALL {0}('{1}',{2},'{3}',{4},'{5}'," \
-              "'{6}','{7}','{8}','{9}',{10}," \
-              "'{11}','{12}','{13}','{14}','{15}'," \
-              "'{16}','{17}','{18}',{19},{20}, SaveStatus)".format(save_sp,
-                                                                   name,
-                                                                   "'{}'".format(model_project)
-                                                                   if model_project else 'NULL',
-                                                                   description,
-                                                                   "'{}'".format(target_column)
-                                                                   if target_column else 'NULL',
-                                                                   prediction_type,
-                                                                   generating_engine,
-                                                                   generating_client,
-                                                                   vantage_version,
-                                                                   algorithm,
-                                                                   "'{}'".format(entity_target)
-                                                                   if entity_target else 'NULL',
-                                                                   build_time,
-                                                                   created_by,
-                                                                   status,
-                                                                   access,
-                                                                   location,
-                                                                   training_objects,
-                                                                   algorithm_parameters,
-                                                                   training_data,
-                                                                   "'{}'".format(performance_metrics)
-                                                                   if performance_metrics else 'NULL',
-                                                                   'NULL')
-
-    # Call the Stored Procedure
-    try:
-        _execute_transaction(sp_call)
-    except SqlOperationalError:
-        raise
-    except Exception as err:
-        raise TeradataMlException(
-            Messages.get_message(MessageCodes.MODEL_CATALOGING_OPERATION_FAILED,
-                                 "save",
-                                 "Stored Procedure call failed."),
-            MessageCodes.MODEL_CATALOGING_OPERATION_FAILED) from err
-
-    # Update Garbage collection, to remove temp table info from the garbage collector.
-    # Print the information for the user experience.
-    print("Persisting model information.")
-    # Let's remove entry for each table from Garbage Collector
-    for table_name in remove_tables_entries_from_gc:
-        dbname = UtilFuncs._teradata_unquote_arg(UtilFuncs._extract_db_name(table_name), "\"")
-        tablename = UtilFuncs._teradata_unquote_arg(UtilFuncs._extract_table_name(table_name), "\"")
-        __gc._delete_object_entry("\"{0}\".\"{1}\"".format(dbname, tablename),
-                                  remove_entry_from_gc_list=True)
-
-        print("Persisted table: {}".format(table_name))
-
-    # Now that we have removed entry from GC, we can safely say that model table has been persisted.
-    print("Successfully persisted model.")
-
-@function_deprecation("future")
-def list_models(name=None,
-                algorithm_name=None,
-                engine=None,
-                accessible=True,
-                public=False):
-    """
-    DESCRIPTION:
-        Function to list models accessible to the user and optionally only models created by the user.
-
-    PARAMETERS:
-        name:
-            Optional Argument.
-            Specifies the search string for model name. When this argument is used,
-            all models matching the name are listed.
-            Types: str
-
-        algorithm_name:
-            Optional Argument.
-            Specifies the search string for the analytic function name. When this argument is used,
-            all models matching the algorithm_name are listed.
-            Types: str
-
-        engine:
-            Optional Argument.
-            Specifies the model generating engine as a filter.
-            Types: str
-            Permitted Values: ['ML Engine', 'Advanced SQL Engine']
-
-        accessible:
-            Optional Argument.
-            Specifies whether to list all models that the user has access to, or only the models that the user created.
-            When True, all models to which the user has access are listed.
-            When False, all models created by the user are listed.
-            Types: bool
-            Default Value: True
-
-        public:
-            Optional Argument.
-            Specifies whether to filter only those models that have public access.
-            Types: bool
-            Default Value: False
-
-    RETURNS:
-        None.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforest", ["housing_train"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-
-        # The examples use home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-        # Let's view the saved models accessible to the user
-        # Example 1 - List all models saved and accessible to the current user.
-        list_models()
-
-        # Example 2 - List models accessible to the user with name = 'decision_forest_model'
-        list_models(name = "decision_forest_model")
-
-        # Example 3 - List all models accessible to the user with algorithm name 'DecisionForest'
-        list_models(algorithm_name = "DecisionForest")
-
-        # Example 4 - List all models accessible to user with algorithm name 'DecisionForest' and model name
-        # containing string 'forest'.
-        list_models(name = "forest", algorithm_name = "DecisionForest")
-
-        # Example 5 - List all models accessible to user with algorithm name 'DecisionForest' and model generated using
-        # 'ML Engine'.
-        list_models(name = "forest", algorithm_name = "DecisionForest", engine = "ML Engine")
-
-        # Example 6 - List all models created by the user with algorithm name 'DecisionForest' and model generated using
-        # 'ML Engine' and access not set to Public.
-        list_models(name = "forest", algorithm_name = "DecisionForest", engine = "ML Engine",
-                    accessible = False, public = False)
-
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, True, (str), True])
-    __arg_info_matrix.append(["algorithm_name", algorithm_name, True, (str), True])
-    __arg_info_matrix.append(["engine", engine, True, (str), True, [mac.MODEL_ENGINE_ML.value,
-                                                                    mac.MODEL_ENGINE_ADVSQL.value]])
-    __arg_info_matrix.append(["public", public, True, (bool)])
-    __arg_info_matrix.append(["accessible", accessible, True, (bool)])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    public = False if not public else True
-    accessible = False if not accessible else True
-
-    # Let's get the required details to list down the models.
-    model_details = DataFrame(in_schema(mac.MODEL_CATALOG_DB.value, mac.MODELS_DETAILSX.value))
-
-    filter_condition = None
-    if not accessible:
-        # We shall show only models created by current user.
-        conn = get_connection()
-        current_user = __get_current_user(conn)
-        filter_condition = model_details[mac.CREATED_BY.value].str.lower() == current_user.lower()
-
-    if public:
-        # Additional filter to filter out models with PUBLIC access.
-        filter_expression = model_details[mac.MODEL_ACCESS.value] == mac.PUBLIC_ACCESS.value
-        if filter_condition:
-            filter_condition = filter_condition & filter_expression
-        else:
-            filter_condition = filter_expression
-
-    if name:
-        # Filter Expression on ModeName column.
-        # We are looking to find all rows with ModelName matching with 'name' string.
-        # This is case-insensitive look-up.
-        filter_expression = __get_like_filter_expression_on_col(model_details._metaexpr,
-                                                                mac.MODEL_DERIVED_NAME.value, name)
-        if filter_condition:
-            filter_condition = filter_condition & filter_expression
-        else:
-            filter_condition = filter_expression
-
-    if algorithm_name:
-        # Filter Expression on ModeAlgorithm column.
-        # We are looking to find all rows with ModeAlgorithm matching with 'algorithm_name' string.
-        # This is case-insensitive look-up.
-        filter_expression = __get_like_filter_expression_on_col(model_details._metaexpr,
-                                                                mac.MODEL_DERIVED_ALGORITHM.value,
-                                                                algorithm_name)
-        if filter_condition:
-            filter_condition = filter_condition & filter_expression
-        else:
-            filter_condition = filter_expression
-
-    if engine is not None:
-        # Filter Expression on ModelGeneratingEngine column.
-        # We are looking to find all rows with ModelGeneratingEngine matching with 'engine' string.
-        # This is case-insensitive look-up.
-        filter_expression = __get_like_filter_expression_on_col(model_details._metaexpr,
-                                                                mac.MODEL_DERIVED_GENENG.value,
-                                                                engine)
-        if filter_condition:
-            filter_condition = filter_condition & filter_expression
-        else:
-            filter_condition = filter_expression
-
-    if filter_condition:
-        model_details = model_details[filter_condition]
-
-    display_df = model_details.select(mac.MODEL_LIST_LIST.value)
-    if display_df.shape[0] != 0:
-        orig_max_rows_num = display.max_rows
-        try:
-            display.max_rows = 99999
-            print(display_df)
-        except Exception:
-            raise
-        finally:
-            display.max_rows = orig_max_rows_num
-    else:
-        # If any of the non-default search-criterion were provided.
-        if name is not None or algorithm_name is not None or engine is not None \
-                or not accessible or public:
-            # MODEL_WITH_SEARCH_CRITERION_NOT_FOUND
-            raise TeradataMlException(
-                Messages.get_message(MessageCodes.MODEL_WITH_SEARCH_CRITERION_NOT_FOUND),
-                MessageCodes.MODEL_WITH_SEARCH_CRITERION_NOT_FOUND)
-        else:
-            # Since no search-criterion were provided by the user,
-            # we print a simple message.
-            print("No models found.")
-
-@function_deprecation("future")
-def describe_model(name):
-    """
-    DESCRIPTION:
-        List details of the model, if accessible to the user.
-
-    PARAMETERS:
-        name:
-            Required Argument.
-            Specifies the name of the model to list the details for.
-            Types: str
-
-    RETURNS:
-        None.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforest", ["housing_train"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-
-        # This example uses home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-        # List all details of recently saved model 'decision_forest_model'.
-        describe_model(name="decision_forest_model")
-
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, False, (str), True])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    # Let's get the required details to list down the models.
-    # We shall show only model accessible by current user.
-    __check_if_model_exists(name, accessible=True, raise_error_if_model_not_found=True)
-
-    df_info_mapper = OrderedDict({mac.MODELS_DETAILS.value: "Model Details",
-                                  mac.MODELS_ATTRS.value: "Model Attributes",
-                                  mac.MODELS_INPUTSX.value: "Model Training Data",
-                                  mac.MODELS_OBJECTS.value: "Model Training Objects",
-                                  mac.MODELS_PERF.value: "Model Performance Metrics",
-                                  mac.MODELS_LOC.value: "Model Location Based Information"
-                                  })
-
-    for info_table, description in df_info_mapper.items():
-        display_df = None
-        if info_table in [mac.MODELS_OBJECTS.value,  mac.MODELS_ATTRS.value, mac.MODELS_INPUTSX.value]:
-            model_client, model_engine, model_algorithm = __retrieve_model_client_engine_algorithm(name)
-            function_arg_map = _argument_mapper._get_function_map(engine=model_engine,
-                                                                  function_name=model_algorithm.lower())
-            if info_table == mac.MODELS_OBJECTS.value:
-                # Output Information.
-                # We wish to display teradataml relevant information, so we get that using this function call
-                # instead of directly reading the underlying view.
-                display_df = __retrieve_model_outputs(name, model_client, function_arg_map)
-            elif info_table == mac.MODELS_ATTRS.value:
-                # Attribute Information.
-                # We wish to display teradataml relevant information, so we get that using this function call
-                # instead of directly reading the underlying view.
-                model_parameters = __retrieve_model_attributes(name, model_client, function_arg_map)[0]
-                attr_name = [key for key in model_parameters.keys()]
-                attr_value = [model_parameters[key] for key in attr_name]
-                display_df = pd.DataFrame({'AttrName': attr_name, 'AttrValue': attr_value})
-            else:
-                # Input data Information.
-                # We wish to display teradataml relevant information, so we get that using this function call
-                # instead of directly reading the underlying view.
-                model_inputs = __retrieve_model_inputs(name, model_client, function_arg_map)
-                input_name = [key for key in model_inputs.keys()]
-                table_name = [model_inputs[key][mac.MODEL_OBJ_TABLE_NAME.value] for key in input_name]
-                nrows = [model_inputs[key][mac.MODEL_INPUT_NROWS.value] for key in input_name]
-                ncols = [model_inputs[key][mac.MODEL_INPUT_NCOLS.value] for key in input_name]
-                display_df = pd.DataFrame({'InputName': input_name,
-                                           'InputTableName': table_name,
-                                           'NRows': nrows,
-                                           'NCols': ncols})
-        else:
-            display_df = DataFrame(in_schema(mac.MODEL_CATALOG_DB.value, info_table))
-            display_df = display_df[display_df[mac.MODEL_DERIVED_NAME.value] == name].to_pandas()
-
-            if info_table == mac.MODELS_DETAILS.value:
-                # We do a transpose of the Model Details to provide better readability of
-                # the ouptut since it is one record with multiple columns.
-                display_df = display_df.set_index(mac.MODEL_DERIVED_NAME.value).transpose()
-            else:
-                # Otherwise, we remove the 'ModelName' column from the
-                # projection which is redundant.
-                cols = display_df.columns.tolist()
-                cols.remove(mac.MODEL_DERIVED_NAME.value)
-                display_df = display_df[cols]
-
-        if display_df.shape[0] > 0:
-            # Set the max_rows parameter to a fairly high value
-            orig_max_rows_num = display.max_rows
-            try:
-                display.max_rows = 99999
-
-                # Print the relevant information for the model
-                print("\n\n*** '{}': {} ***".format(name, description))
-                print(display_df)
-            except Exception:
-                raise
-            finally:
-                # Reset the max_rows parameter to it's original value
-                display.max_rows = orig_max_rows_num
-
-@function_deprecation("future")
-def delete_model(name, delete_objects=False):
-    """
-    DESCRIPTION:
-        Delete a model, and optionally delete the model objects.
-        A model can be deleted only by the creator of the model.
-
-    PARAMETERS:
-        name:
-            Required Argument.
-            Specifies the name of the model to be deleted.
-            Types: str
-
-        delete_objects:
-            Optional Argument.
-            Specifies whether to drop the model objects also.
-            When True, the model objects related to the model are deleted/dropped.
-            Types: bool
-            Default Value: False
-
-    RETURNS:
-        None.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforest", ["housing_train"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-
-        # The examples use home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-        # Example 1 - Only delete model information from the Model Catalog.
-        delete_model('decision_forest_model')
-
-        # Example 2 - Delete model information from the Model Catalog and drop model objects as well.
-        delete_model('decision_forest_model', True)
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, False, (str), True])
-    __arg_info_matrix.append(["delete_objects", delete_objects, True, (bool)])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    # Let's check if the user created the model since only the creator can delete it
-    __check_if_model_exists(name, created=True, raise_error_if_model_not_found=True)
-
-    conn = get_connection()
-    current_user = __get_current_user(conn)
-
-    # Get the models tables for the given model
-    model_tables = __get_tables_for_model(name, current_user)
-
-    # First, find the transaction mode of the session to decide which Stored Procedure to call
-    delete_sp = '{}{}'.format(mac.DELETE_MODEL.value, 'T')
-    transaction_mode = conn.dialect.get_transaction_mode(conn)
-    if transaction_mode and transaction_mode.lower() == 'a':
-        delete_sp = '{}{}'.format(mac.DELETE_MODEL.value, 'A')
-
-    # Form the stored procedure call
-    sp_call = "CALL {0}('{1}', DeleteStatus)".format(delete_sp, name)
-
-    # Call the stored Procedure after any grant statements and only if they were successful
-    try:
-        # Call the Stored Procedure
-        _execute_transaction(sp_call)
-    except SqlOperationalError:
-        raise
-    except Exception as ex:
-        raise TeradataMlException(
-            Messages.get_message(MessageCodes.MODEL_CATALOGING_OPERATION_FAILED,
-                                 "delete",
-                                 "Stored Procedure call failed."),
-            MessageCodes.MODEL_CATALOGING_OPERATION_FAILED) from ex
-
-    print("Deleted model '{}' successfully.".format(name))
-
-    if delete_objects:
-        failed_to_drop = []
-        for table_name in model_tables:
-            if not __delete_model_tableview(table_name):
-                failed_to_drop.append(table_name)
-        if len(failed_to_drop) > 0:
-            warnings.warn('Failed to drop the following Model Objects: {}.'.format(failed_to_drop))
-        else:
-            print("Model Objects dropped successfully.")
-    else:
-        print('Model Objects that can be dropped: {}.'.format(model_tables))
-
-@function_deprecation("future")
-def retrieve_model(name):
-    """
-    DESCRIPTION:
-        Function to retrieve a saved model accessible to the current user.
-
-    PARAMETERS:
-        name:
-            Required Argument.
-            Specifies the name of the model to retrieve.
-            Types: str
-
-    RETURNS:
-        The Analytic function object corresponding to the model retrieved.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforestpredict", ["housing_train","housing_test"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-        housing_test = DataFrame.from_table("housing_test")
-
-        # This example uses home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-        # Retrieve the saved model.
-        retrieved_rft_model = retrieve_model("decision_forest_model")
-
-        # Use the retrieved model in predict.
-        decision_forest_predict_out = DecisionForestPredict(object = retrieved_rft_model,
-                                                            newdata = housing_test,
-                                                            id_column = "sn",
-                                                            detailed = False,
-                                                            terms = ["homestyle"],
-                                                            newdata_order_column=['sn', 'price'],
-                                                            object_order_column=['worker_ip', 'task_index']
-                                                           )
-
-        # Print the results.
-        decision_forest_predict_out.result
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, False, (str), True])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    # We shall show only retrieve models accessible by current user.
-    # Let's first make sure, whether such model with name exists or not.
-    __check_if_model_exists(name, accessible=True, raise_error_if_model_not_found=True)
-
-    # Let's read model attributes and their values/output tables.
-    model_class, model_engine, attr_and_output_dictionary = __retrieve_argument_and_output_map(name)
-
-    # Get wrapper class for which model is to be instantiated. We shall later use the same to instantiate the
-    # wrapper.
-    class_ = __get_wrapper_class(model_engine, model_class)
-
-    # Let's instantiate model class and return the model of CLASS model_class.
-    # But first we make sure the keys are strings
-    new_dict = {}
-    for key in attr_and_output_dictionary:
-        val = attr_and_output_dictionary[key]
-        if not isinstance(key, str):
-            key = str(key)
-
-        new_dict[key] = val
-
-    return class_._from_model_catalog(**new_dict)
-
-
-# TODO: For publish model
-# 1. Should have a way to issue GRANTs/REVOKEs
-# 2. Should also have a way to undo the publish operation (revoke access) and set access to Private
-#    or a different role/user.
-#    Need to get a list of users who already have access to it
-@function_deprecation("future")
-def publish_model(name, grantee=None, status=None):
-    """
-    DESCRIPTION:
-        Function to publish a model.
-        It can be used to update the access level from PRIVATE to PUBLIC or TEAM.
-        It can also be used to update the status of the model to one of the expected values.
-        A model can be published only by the creator of the model.
-
-    PARAMETERS:
-        name:
-            Required Argument.
-            Specifies the name of the model to be published.
-            Types: str
-
-        grantee:
-            Optional Argument. Must be specified when status is not specified.
-            Specifies a user or role (including PUBLIC) to update the model's access level to.
-            Types: str
-
-        status:
-            Optional Argument. Must be specified when grantee is not specified.
-            Specify a string to set the status of the model to.
-            Types: str
-            Permitted Values: ['ACTIVE', 'RETIRED', 'CANDIDATE', 'PRODUCTION', 'IN-DEVELOPMENT']
-
-    RETURNS:
-        None.
-
-    RAISES:
-        TeradataMlException, TypeError, ValueError
-
-    EXAMPLES:
-        # Load the data to run the example
-        load_example_data("decisionforest", ["housing_train"])
-
-        # Create teradataml DataFrame objects.
-        housing_train = DataFrame.from_table("housing_train")
-
-        # The examples use home sales data to create a
-        # classification tree that predicts home style, which can be input
-        # to the DecisionForestPredict.
-        formula = "homestyle ~ driveway + recroom + fullbase + gashw + airco + prefarea + price + lotsize + bedrooms + bathrms + stories + garagepl"
-        rft_model = DecisionForest(data=housing_train,
-                                   formula = formula,
-                                   tree_type="classification",
-                                   ntree=50,
-                                   tree_size=100,
-                                   nodesize=1,
-                                   variance=0.0,
-                                   max_depth=12,
-                                   maxnum_categorical=20,
-                                   mtry=3,
-                                   mtry_seed=100,
-                                   seed=100
-                                   )
-
-        # Let's save this generated model.
-        save_model(model=rft_model, name="decision_forest_model", description="Decision Forest test")
-
-        # Example 1 - Update only the access.
-        publish_model('decision_forest_model', grantee='john')
-
-        # Example 2 - Update only the status.
-        publish_model('decision_forest_model', status='Active')
-
-        # Example 3 - Update both the access and status.
-        publish_model('decision_forest_model', grantee='PUBLIC', status='Production')
-    """
-    # Perform required validations for the API.
-    # Check whether Model Cataloging tables exist or not.
-    __check_if_model_cataloging_tables_exists()
-
-    # Let's perform argument validations.
-    # Create argument information matrix to do parameter checking
-    __arg_info_matrix = []
-    __arg_info_matrix.append(["name", name, False, (str), True])
-    __arg_info_matrix.append(["grantee", grantee, True, (str), True])
-    __arg_info_matrix.append(["status", status, True, (str), True, mac.MODEL_VALID_STATUS.value])
-
-    # Make sure that a correct type of values has been supplied to the arguments.
-    __awu._validate_function_arguments(__arg_info_matrix)
-
-    # Both grantee and status cannot be null
-    if grantee is None and status is None:
-        raise TeradataMlException(Messages.get_message(MessageCodes.EITHER_THIS_OR_THAT_ARGUMENT, "grantee", "status"),
-                                  MessageCodes.EITHER_THIS_OR_THAT_ARGUMENT)
-
-    # Let's check if the user created the model since only the creator can publish it
-    __check_if_model_exists(name, created=True, raise_error_if_model_not_found=True)
-
-    conn = get_connection()
-
-    # First, find the transaction mode of the session to decide which Stored Procedure to call
-    publish_sp = '{}{}'.format(mac.PUBLISH_MODEL.value, 'T')
-    transaction_mode = conn.dialect.get_transaction_mode(conn)
-    if transaction_mode and transaction_mode.lower() == 'a':
-        publish_sp = '{}{}'.format(mac.PUBLISH_MODEL.value, 'A')
-
-    print_message = None
-    current_user = __get_current_user(conn)
-    old_access = __get_model_access(name)
-    # Update the access only when we have a grantee that is not the same as the current user or 'private',
-    # and when the current access level is PRIVATE.
-    if grantee is not None and grantee.lower() != current_user.lower() \
-            and grantee.lower() != 'private' and old_access.lower() == 'private':
-        grant_list = __get_tables_for_model(name, current_user)
-
-        print_message = "Please execute the following GRANT statements:\n{}".format(
-            '\n'.join('GRANT SELECT ON {} to {};'.format(obj, grantee) for obj in grant_list))
-    elif grantee is not None:
-        warnings.warn("'grantee' must not be equal '{0}' or the current user's name,"
-                      " and the earlier access level must be '{0}'.".format(mac.DEFAULT_SAVE_ACCESS.value))
-        # Set grantee to None so that it is not passed in stored procedure
-        grantee = None
-
-    # Call the stored procedure after any grant statements and only if they were successful
-    # Form the stored procedure call
-    sp_call = "CALL {0}('{1}',{2},{3}, PublishStatus)".format(publish_sp,
-                                                              name,
-                                                              "'{}'".format(grantee) if grantee else 'NULL',
-                                                              "'{}'".format(status) if status else 'NULL'
-                                                              )
-
-    try:
-        # Call the Stored Procedure
-        _execute_transaction(sp_call)
-    except SqlOperationalError:
-        raise
-    except Exception as ex:
-        raise TeradataMlException(
-            Messages.get_message(MessageCodes.MODEL_CATALOGING_OPERATION_FAILED,
-                                 "publish",
-                                 "Stored Procedure call failed."),
-            MessageCodes.MODEL_CATALOGING_OPERATION_FAILED) from ex
-
-    print("Model published successfully!")
-    if print_message:
-        print(print_message)